Polygenic risk score values vary with genetic ancestry due to differences in population-specific allele frequencies and linkage disequilibrium patterns. We present a framework to calibrate polygenic risk scores based on ancestral makeup. We propose the "expected polygenic risk score" or ePRS, defined as the expected value of a polygenic risk score based on one's global or local admixture patterns. We further define the "residual polygenic risk score" or rPRS as measuring the deviation of the polygenic risk score from the ePRS. The ePRS reflects the baseline ancestry-driven component of genetic risk, whereas the rPRS isolates an ancestry-agnostic measure of genetic liability. Simulation studies confirm that it suffices to adjust for ePRS to obtain nearly unbiased estimates of the polygenic risk score-outcome association without further adjusting for principal components. Using the TOPMed and the All of Us datasets, effect size estimates for the rPRS (adjusted for ePRS) are similar to those obtained from polygenic risk scores adjusting for genetic principal components. The ePRS framework can protect from population stratification in association analysis and provide an equitable strategy to interpret genetic risk across diverse populations.
Polygenic Risk Scores
Prediction models for cognitive aging measures have largely evaluated demographic variables and APOE carrier status in populations of European ancestry. To comprehensively assess prediction models among Hispanic/Latinos, we considered 12 models (6 predictor sets and 2 methods) for global cognitive score change (GCSC) and mild cognitive impairment (MCI) in the Study of Latinos-Investigation of Neurocognitive Aging (SOL-INCA) (N = 5856). Based on the average mean squared error (MSE) for GCSC or average area under the curve (AUC) for MCI across 100 randomly split testing and training sets, performance was similar across models, but slightly better for the following models: the chronic conditions model and genetic model (mean MSEs = 0.2464) using gradient-boosted trees for GCSC prediction and the chronic conditions model (mean AUC = 62%) and metabolite model (mean AUC = 60%) using logistic regression for MCI prediction. Using the Shapley Additive Explanations (SHAP) method, age at baseline, time between exams, and sex were the most important predictors for GCSC, followed by diabetes and global ancestral proportions. Diabetes and the metabolite ribitol had the highest influence on prediction of MCI. Although prediction performance was not especially high and did not vary greatly across models, incorporating information on diabetes, ancestry and metabolites may help improve prediction of GCSC and MCI.
Over 75 Alzheimer's disease (AD) and dementia-associated variants have been identified through genome-wide association studies, but the utility of polygenic risk scores (PRS) for predicting AD and dementia in diverse and admixed populations remains unclear. We compared how PRS approaches differing in p-value thresholds, variant weights, and source ancestry perform in predicting dementia in 6338 African American, Chinese, Hispanic, and White individuals from the Multi-Ethnic Study of Atherosclerosis. We tested clumping and thresholding (C+T) methods with varying parameters against Bayesian approaches (PRS-CS, PRS-CSx). We compared the ability of each method to predict incident dementia in all participants and in groups stratified by self-reported race/ethnicity. We additionally analyzed performance across groups stratified by estimated proportion of non-Finnish European (NFE)-like ancestry. Including more variants does not improve performance. We found comparable associations between dementia and PRS when comparing a C+T method with only 15 SNPs and PRS derived from Bayesian models that include > 800,000 SNPs (HR5e-08 = 1.18, 95% CI: 1.08-1.28; HRCSx = 1.17, 95% CI: 1.07-1.27). The p < 5e-08 C+T method was more strongly associated with incident dementia in populations genetically dissimilar from the source data (HRlowNFE_5e-08 = 1.27, 95% CI: 1.08-1.50; HRlowNFE_CSx = 1.12, 95% CI: 0.94-1.33). More selective PRS models using genome-wide significant SNPs may be preferable for dementia prediction in diverse populations.
Polygenic scores (PGSs) offer moderate to high prediction accuracy for complex traits, but most are developed in European ancestry cohorts, reducing their performance in populations of other ancestries. This study aimed to improve standing height prediction, a heritable and ancestry-influenced trait, in an admixed Latino cohort, the Hispanic Community Health Study/Study of Latinos (HCHS/SOL), by modeling ancestry using principal components (PCs) alongside PGSs. SNPs were selected from a large European ancestry genome-wide association study (GWAS) using various p value thresholds, and weights were trained using traditional and penalized regression in the UK Biobank (UKB). PGSs with PCs were trained separately in the HCHS/SOL and UKB. Compared to PGSs alone, modeling PGSs with PCs moderately improved height prediction in the HCHS/SOL (squared correlation [R2] increase of ∼0.05), while mild improvements were observed in the UKB (R2 increase of ∼0.01). These results underscore the importance of incorporating genetic ancestry into predictive models for admixed populations, particularly when the trait exhibits ancestry-specific associations.
STUDY OBJECTIVES: Since genome-wide association studies (GWAS) of sleep phenotypes have been conducted in differing populations and definitions of sleep phenotypes vary across studies, we investigated associations between several polygenic risk scores (PRSs) and potential sleep definitions among multiethnic cohorts.
METHODS: Using data from four cohorts (HCHS/SOL, ARIC, MESA, BHS, N = 16 895), we considered multiple definitions of short and long sleep, insomnia, and excessive daytime sleepiness (EDS). PRSs were developed based on summary statistics from GWAS in European ancestry individuals from the UK Biobank (UKB) and from GWAS conducted in a multiethnic population from the Million Veteran Program (MVP). Study-specific analyses estimated associations between sleep PRSs and corresponding sleep measures per 1 standard deviation increase in the PRS. Models were adjusted for age, sex, ancestral principal components, and center and race as appropriate. Results were meta-analyzed across studies.
RESULTS: PRSs based on European ancestry UKB GWAS had statistically significant associations with multiple definitions of the corresponding sleep phenotypes. Associations that were most consistent across studies included: short sleep PRS with ≤6 hours (OR = 1.23,p = 1.80x10-7,phet = 0.97); long sleep PRS with ≥9 hours (OR = 1.09,p = 6.76x10-4,phet = 0.77); insomnia PRS with the Women's Health Initiative Insomnia Rating Scale (WHIIRS) ≥10 or a subset of three questions ≥6 in ARIC (OR = 1.17,p = 5.51x10-10,phet = 0.69); and EDS PRS with Epworth Sleepiness Scale (ESS) ≥11 (OR = 1.23,p = 1.83x10-13,phet = 0.83). PRSs based on multi-ancestry MVP GWAS had weaker associations compared to those based on European ancestry only.
CONCLUSIONS: By evaluating several types of sleep PRSs and sleep phenotypes, we were able to highlight which sleep PRS performed well across diverse populations and which sleep definitions better captured genetic underpinnings.
In a multi-stage analysis of 52,436 individuals aged 17-90 across diverse cohorts and biobanks, we train, test, and evaluate a polygenic risk score (PRS) for hypertension risk and progression. The PRS is trained using genome-wide association studies (GWAS) for systolic, diastolic blood pressure, and hypertension, respectively. For each trait, PRS is selected by optimizing the coefficient of variation (CV) across estimated effect sizes from multiple potential PRS using the same GWAS, after which the 3 trait-specific PRSs are combined via an unweighted sum called "PRSsum", forming the HTN-PRS. The HTN-PRS is associated with both prevalent and incident hypertension at 4-6 years of follow up. This association is further confirmed in age-stratified analysis. In an independent biobank of 40,201 individuals, the HTN-PRS is confirmed to be predictive of increased risk for coronary artery disease, ischemic stroke, type 2 diabetes, and chronic kidney disease.
Polygenic risk scores (PRS) are commonly used to quantify the inherited susceptibility for a trait, yet they fail to account for non-linear and interaction effects between single nucleotide polymorphisms (SNPs). We address this via a machine learning approach, validated in nine complex phenotypes in a multi-ancestry population. We use an ensemble method of SNP selection followed by gradient boosted trees (XGBoost) to allow for non-linearities and interaction effects. We compare our results to the standard, linear PRS model developed using PRSice, LDpred2, and lassosum2. Combining a PRS as a feature in an XGBoost model results in a relative increase in the percentage variance explained compared to the standard linear PRS model by 22% for height, 27% for HDL cholesterol, 43% for body mass index, 50% for sleep duration, 58% for systolic blood pressure, 64% for total cholesterol, 66% for triglycerides, 77% for LDL cholesterol, and 100% for diastolic blood pressure. Multi-ancestry trained models perform similarly to specific racial/ethnic group trained models and are consistently superior to the standard linear PRS models. This work demonstrates an effective method to account for non-linearities and interaction effects in genetics-based prediction models.
Estimated glomerular filtration rate (eGFR) is used to evaluate kidney function and determine the presence of chronic kidney disease (CKD), a highly prevalent disease in the US1 , 2 , 3 that varies among subgroups of Hispanic/Latino individuals.4 , 5 The polygenic risk score (PRS) is a popular method that uses large genome-wide association studies (GWASs) to provide a strong estimate of disease risk.7 However, due to the limited availability of summary statistics from GWAS meta-analyses based on Hispanic/Latino populations, PRSs can only be computed using different ancestry GWASs. The performance of eGFR PRSs derived from other GWAS reference populations for Hispanic/Latino population has not been examined. We compared PRS constructions for eGFR prediction in Hispanic/Latino individuals using GWAS-significant variants, clumping and thresholding (C&T),8 and PRS-CS,22 as well as a combination of PRSs calculated with different reference GWAS meta-analyses from European and multi-ethnic studies in Hispanic/Latino individuals from the Hispanic Community Health Study/Study of Latinos (HCHS/SOL). All eGFR PRSs were highly associated with eGFR (p < 1E-20). Additionally, eGFR PRSs were significantly associated with lower risk of prevalent CKD at visit 1 or 2 and incident CKD at visit 2, with the combined PRSs having the best performance. These PRS findings were replicated in an additional dataset of Hispanic/Latino individuals using data from the Women's Health Initiative SNP Health Association Resource (WHI-SHARe).17.
OBJECTIVE: In the United States, Hispanic/Latino adults face a high burden of obesity; yet, not all individuals are equally affected, partly due in part to this ethnic group's marked sociocultural diversity. We sought to analyze the modification of body mass index (BMI) genetic effects in Hispanic/Latino adults by their level of acculturation, a complex biosocial phenomenon that remains understudied.
METHODS: Among 11,747 Hispanic/Latinos adults in the Hispanic Community Health Study/Study of Latinos aged 18 to 76 years from four urban communities (2008-2011), we a) tested our hypothesis that the effect of a genetic risk score (GRS) for increased BMI may be exacerbated by higher levels of acculturation and b) examined if GRS acculturation interactions varied by gender or Hispanic/Latino background group. All genetic modeling controlled for relatedness, age, gender, principal components of ancestry, center, and complex study design within a generalized estimated equation framework.
RESULTS: We observed a GRS increase of 0.34 kg/m 2 per risk allele in weighted mean BMI. The estimated main effect of GRS on BMI varied both across acculturation level and across gender. The difference between high and low acculturation ranged from 0.03 to 0.23 kg/m 2 per risk allele, but varied across acculturation measure and gender.
CONCLUSIONS: These results suggest the presence of effect modification by acculturation, with stronger effects on BMI among highly acculturated individuals and female immigrants. Future studies of obesity in the Hispanic/Latino community should account for sociocultural environments and consider their intersection with gender to better target obesity interventions.
Polygenic risk scores (PRSs) are weighted sums of risk allele counts of single-nucleotide polymorphisms (SNPs) associated with a disease or trait. PRSs are typically constructed based on published results from Genome-Wide Association Studies (GWASs), and the majority of which has been performed in large populations of European ancestry (EA) individuals. Although many genotype-trait associations have generalized across populations, the optimal choice of SNPs and weights for PRSs may differ between populations due to different linkage disequilibrium (LD) and allele frequency patterns. We compare various approaches for PRS construction, using GWAS results from both large EA studies and a smaller study in Hispanics/Latinos: The Hispanic Community Health Study/Study of Latinos (HCHS/SOL, n = 12 , 803 ). We consider multiple approaches for selecting SNPs and for computing SNP weights. We study the performance of the resulting PRSs in an independent study of Hispanics/Latinos from the Women's Health Initiative (WHI, n = 3 , 582 ). We support our investigation with simulation studies of potential genetic architectures in a single locus. We observed that selecting variants based on EA GWASs generally performs well, except for blood pressure trait. However, the use of EA GWASs for weight estimation was suboptimal. Using non-EA GWAS results to estimate weights improved results.
